The increase in numbers and contamination levels of faecally polluted water has resulted in shifts worldwide towards methods which enumerate faecal indicator bacteria faster. Rapid methods enable more timely remedial and preventative actions which protect the health of water users. However, especially in the developing world, straightforward methods are also preferred as they reduce the requirement for highly qualified analysts. This study investigates the feasibility of using the rapid, semi-automated enzyme substrate test Colilert-18® instead of multiple-tube fermentation (MTF) in total coliform and Escherichia coli enumeration for South African river water, as one example of a surface water source carrying considerable faecal pollution, which needs monitoring. Spearman rank correlation coefficients (ρ) of 0.83 and 0.86 were obtained for total coliforms and E. coli respectively, indicating Colilert-18® performed acceptably in the pollution ranges encountered. A Bland–Altman plot further revealed that Colilert-18® showed no significant difference (p > 0.05) from MTF values below 100,000 E. coli most probable number/100 mL (estimated true value). Above this level Colilert-18® was found to progressively underestimate E. coli. This inadequacy of Colilert-18® was considered acceptable from a health risk assessment viewpoint as such high counts should have sounded the alarm for preventative and corrective action irrespective of method inaccuracy.

You do not currently have access to this content.